Beating AI News Flash: Anthropic has released an economic study simulating how AI might affect the U.S. economy by 2030. The study first sets assumptions about how much work AI can do and how quickly businesses will adopt it, then calculates the corresponding GDP, wages, and unemployment rate.
In the most aggressive scenario, AI would affect about 30% of work tasks across the entire economy. U.S. GDP would be 32.4% higher than without AI, but the unemployment rate for knowledge workers would rise to 17.9%, and the overall unemployment rate would reach 11.9%. Knowledge workers' wages would also be 11.5% lower.
Changes in income distribution would also be striking. Labor's share of income would fall from 60% to 45.2%, while capital's share would rise to 54.8%. Although the economy would expand by nearly one-third, total labor income would rise by only 0.5%, while capital income would rise by 81.4%.
But under a less aggressive set of assumptions, the results would be much more moderate. The typical response in U.S. public surveys is closer to the middle scenario, in which GDP by 2030 would be about 10% higher than without AI, and the overall unemployment rate would be about 5%.

